Re: [CAD_CAM_EDM_DRO] DRO

>
> From: beer@...
>
> > Thanks James,
> >
> > I knew about US Digital. I was hoping this was something
> > else, as I was looking for a little higher resolution
> > linear encoders. (cheap)
> >
>
>
> The USDigital strips CAN be used at much higher resolutions.
>
> The trick here is to NOT use the HP "digital" decoders, but to use an
> analog optical decoder ( of some custom design ).
>
> Instead of sensing "fully black" or "fully white", one would
> additionally sense intermediate light levels, e.g., "half white". The
> number of intermediate levels available would depend on the edge
> sharpness of the lines on the strip, but doubling the resolution of the
> strips should present no real challenge and even tripling or
> quadrupling it could be possible.
>
> Some time would have to be spent aligning the two sensors and for this,
> a dual trace scope makes life a lot easier. But alignment presents no
> real difficulty, just tedium.
>
> Ultimately, diffraction effects would start to rear its ugly head to
> limit the information available via this simple means. Also, it might
> be difficult to set up so that each "step" occurs at exactly the same
> distance at higher "multiples" than just a doubling.
>
> The thing is, it IS cheap, as you won't be buying the HP parts,
> you'll be making your own.

I don't know exactly how these strips are set up,
I don't recognize the pattern printed on them, or the method of
reading them.
If if this ends up being and optical do-it-yourself kind of thing,
I looked into having a local typesetter print out a strip of film.

His linotype machine could print at about 2400 dpi If I recall.
It seems like he Mentioned something like 7,000 being possible
with his new machine.

It sounds like pretty much the same film and process used for the US Digital
strips and disks.

I wondered about a vernier type setup, Think quadrature then keep moving
the sensors to split the previous pulse.

I also wondered about an absolute position Gray Code strip using 16 bits
or so.

As to adjustment, I figured on using a logic analyzer.
I have one good to 50 Mhz, and 48 inputs. Since in the end the outputs
from the system I figured on using was digital, ie. light OR dark, this
would work great, and allow all the outputs to be tracked at the same
instant.

I figured on printing the shadow masks for the sensors using the same
process as the strip itself, as that negates the need for drilling holes
or slots to the very high tolerances required (Relative to any machine
I have).


The question I have with either type of strip is how to mount it, and
how to both allow the reader to move across the surface and keep the
dirt, oil, coolant, and chips out?
One chip stuck in the sensor gap and scraped down the surface would
ruin it.

Marc